Top news of the week: 15.06.2022.
Learning
Machine learning
Recurrence relation
Human
Psychology
Ontology
RT @Yuhu_ai_: We open sourced Memorizing Transformers (https://t.co/g7xItsTffv) and Block Recurrent Transformers (https://t.co/mogGALzGQV) in Meliad!
Repo link: https://t.co/GhoLHyieLA https://t.co/3hDJ8x4pAY
Opengoogle-research/meliad
Contribute to google-research/meliad development by creating an account on GitHub.
@jsotterbach @katecrawford Our mutual chum @Hernandez_Danny did this analysis based on the data from this paper https://t.co/ZfbnUdqlgu, iirc.
Open
Compute Trends Across Three Eras of Machine Learning
Compute, data, and algorithmic advances are the three fundamental factors that guide the progress of modern Machine Learning (ML). In this paper we study trends in the most readily ...
RT @drob: . @sarahcat21 is revisiting the classic AI Hierarchy of Needs https://t.co/oJ0NoHKqmd
Includes some 🔥 about how the modern data stack is graded on each level 😬 #rstatsnyc https://t.co/A3knw3xEG6
Open
The AI Hierarchy of Needs
But the most common scenario is that they have not yet built the infrastructure to implement (and reap the benefits of) the most basic data science algorithms and operations, much …
For people, identity is forever in flux. To a machine, human identity cannot evolve, iterate or be subverted.
This work, forthcoming at @FAccTConference ’22, critiques the existing paradigm of identity in AI. By @t1ngyu3, @jackayline & @empiricallykev.
https://t.co/33LQLX0aTd https://t.co/7QVR1wC7ZC
OpenSubverting machines, fluctuating identities: Re-learning human categorization
Most machine learning systems that interact with humans construct some notion of a person's "identity," yet the default paradigm in AI research envisions identity with essential attributes ...